Generalized Estimating Equations ( GEE )

Abstract Correlated datasets arise from repeated measures studies where multiple observations are collected from a specific sampling unit (a specific patient's status over time), or from grouped orclustered datawhere observations are grouped based on sharing some common characteristic (animals in a specific litter). When measurements are collected over time, the term longitudinal or panel data is preferred. Generalized estimating equations provide a framework for analyzing correlated data. This framework extends the generalized linear models methodology, which assumes independent data. We discuss the estimation of model parameters and associated variances via generalized estimating equation methodology.